Differential abundance testing in microbiome data poses significant challenges due to its sparse, highly variable, and compositional nature. The application of statistical methods, both parametric and non-parametric, in this context is complex. Currently, many statistical methods in this field assume classical distribution models or take into account compositional specifics, resulting in a trade-off between specificity and sensitivity. This makes it difficult to assess type I and type II errors when a single method is used. To address this challenge, a consensus approach based on multiple differential abundance (DA) methods has been proposed to increase robustness.

The ideal candidate for this master's project should have knowledge of programming in R, as well as experience with Git and GitHub. This expertise will be valuable in contributing to the development of an R package that aims to simplify and enhance the process of differential abundance testing in microbiome data. The package allows users to use pipeable sequences of AD methods, similar to dplyr syntax, and apply various consensus strategies. This approach has the potential to produce more reliable, consistent, and reproducible results in microbiome data analysis. 

Furthermore, there is an opportunity for the ideal student to continue their research in the laboratory and explore the possibility of pursuing a Ph.D. in the field of computational bioinformatics.

Supervisors: Francesc Català, Roger Paredes